FFmpegFrameGrabber's timeout for RTSP #711

[d-a-gerashenko]

Two methods to test "connection is lost" situations. In addition to #711 issue.

[saudet]

Thanks for the contribution! Since this works only for RTSP, maybe we could name the sample RtspStreamingTimeout.java, what do you think?

[d-a-gerashenko]

Ok. Should I do it or you will rename it?

[saudet]

Yes, please rename! Thanks

[saudet]

It looks like we can set the "timeout" for the FTP, HTTP, RTMP, SMB, SSH, TCP, UDP, and UNIX protocols as well, in addition to a generic "rw_timeout": http://ffmpeg.org/ffmpeg-all.html#Protocols

Would you mind adding this in commented lines of (even if untested) code?
And then we can name the file, maybe, "FFmpegStreamingTimeout.java"? :)

[d-a-gerashenko]

Interesting. Yes, sure) I'll test it. If it works there is no need in "stimeout".

[d-a-gerashenko]

Nope. "rw_timeout" isn't working. Actually it's working when grabber is already connected and has lost connection. But it's not working when network interface was disabled before connection attempt.

[saudet]

Ok, cool, thanks for testing. Let's just leave that as well as "timeout" for the other protocols in comment.

[d-a-gerashenko]

continuous-integration/travis-ci/pr — The Travis CI build failed

Is it normal?

I didn't change the file's name because the option doesn't work the way we expected. Right?

[saudet]

Yeah, I still need to get Travis going. You can ignore that for now.

[saudet]

I'm sure the "timeout" and "rw_timeout" options work for some protocols...

[d-a-gerashenko]

Ok. I can rename it to FFmpegStreamingTimeout but I will add comment that it was tested with RTSP only.

[saudet]

Sounds good! Thanks

[d-a-gerashenko]

I think it's done.

[saudet]

Thanks! Had to fix the filename though

[d-a-gerashenko]

Oops. Sorry)